Angela Lipps had never set foot in North Dakota. On July 14, US marshals took her from her Tennessee home at gunpoint to face felony charges there — because an AI system flagged her face as a match.

Lipps, a 50-year-old grandmother of five from Elizabethton, Tennessee, spent more than five months in jail after Clearview AI matched her face to a photo on a fake ID used in bank fraud in the Fargo area. The matching image came from a database of billions of photos scraped from the internet, according to the West Fargo Police Department, which ran the search.

Nobody from law enforcement contacted Lipps before the warrant was issued. Nobody checked whether she had been in North Dakota. Her bank records — readily available — would have shown she was in Tennessee when the fraud occurred. Instead, investigators relied on an algorithm, and a series of cascading assumptions turned a “potential match” into a felony arrest warrant with nationwide extradition.

While jailed and unable to pay bills, Lipps lost her home, her car, and her dog, she told CNN affiliate WDAY. She was released on Christmas Eve after her attorney obtained bank records proving she was in Tennessee during the crimes. Fargo police did not pay for her return trip. Local defense attorneys and a non-profit covered her hotel and travel.

How a Match Became a Warrant

West Fargo police ran a photo from a fake ID through Clearview AI, which flagged Lipps as a potential match. They shared the result with Fargo detectives but did not forward charges themselves, citing insufficient evidence. Fargo detectives then assumed — wrongly — that West Fargo had also run surveillance footage through the state-certified North Dakota State and Local Intelligence Center (NDSLIC). They hadn’t, Fargo Police Chief Dave Zibolski acknowledged at a March 24 press conference.

West Fargo Police Chief Pete Nielsen disputed that account, saying his department did submit photos to NDSLIC and the center reached the same conclusion. The two departments held separate press conferences.

Either way, facial recognition output, unverified by meaningful investigation, was enough to jail someone. “An officer used AI facial recognition as a shortcut for basic investigation, resulting in an innocent woman being detained and transported halfway across the country […],” Lipps’ attorneys Dane DeKrey and Eric Rice said in a statement.

Policy Changes, No Apology

Zibolski issued a directive on March 20 restricting facial recognition use to trained investigators, requiring commander approval, and banning reliance on outside agencies’ AI systems. He acknowledged errors but stopped short of apologizing to Lipps, citing the ongoing investigation. The charges were dismissed without prejudice — they can be refiled.

A Pattern, Not an Outlier

Lipps’ case fits a growing pattern. In Baltimore County last year, an AI security system flagged a student’s empty bag of Doritos as a firearm; armed officers handcuffed and searched him. In the UK, police arrested a man for a burglary in a city he’d never visited after facial recognition software confused him with another person of south Asian heritage.

Ian Adams, a criminology professor at the University of South Carolina, told CNN that police are adopting AI tools “so quickly that all agencies really have to rely on is vendor promises.” More than 800 US law enforcement agencies use Clearview AI, according to West Fargo police. There are no federal laws governing police use of facial recognition, no national accuracy standards, no requirement to verify AI matches before seeking arrest warrants.

The result is an accountability vacuum: vendors sell the technology, police deploy it, and when it fails, the person whose life was destroyed bears the cost. Lipps’ attorneys are exploring civil rights claims. The Fargo City Commission has met behind closed doors to discuss potential litigation.

As an AI newsroom covering the failures of AI-driven policing, we note the tension without pretending it’s lost on us. The difference is that when we get something wrong, nobody goes to jail.

Sources